Prognostic factors in primary breast carcinoma.

The range of therapeutic options for the treatment of patients with primary breast cancer has widened considerably in recent years. The introduction of mammographic screening has led to the identification of greater numbers of small, stage 1 cancers. It is therefore increasingly important that clinicians are given the most accurate prognostic information on which to base the selection of the optimum therapy for each woman. In this way both under and over treatment can be avoided. This clinical need for information has resulted in an ever widening search for the ideal prognostic factor in primary breast cancer and consequently a multitude of papers have proposed numerous putative prognostic factors with differing degrees of significance. The current obsession with high technology has, to some extent, led to the value of traditional histopathology being overlooked, although the correlation between tumour morphology and degree of malignancy was first described over 100 years ago. This is unfortunate because routine histopathological examination of breast cancer specimens can yield a great deal of clinically useful information. Histological grade provides prognostic information in many tumours, including breast cancers. Two main methods have evolved, based either on nuclear factors or a combination of cellular features (nuclear, cytological and architectural). The latter method is the most widely accepted for grading breast cancer and has been refined with the stricter definition of more objective criteria.' As a result, the reproducibility of histological grading systems for breast carcinoma, previously questioned, has been improved; several recent studies have confirmed that good correlation between pathologists can be obtained if strict criteria are used.2" In the Nottingham breast cancer grading method the three features assessed are tubule formation, nuclear size/pleomorphism and mitotic count. Tumours in which less than 10% of the carcinoma is forming tubules score 3, those with 10 to 75% score 2 and if more than 75% of the cancer is forming central lumina, a score of 1 is given. Mild, moderate and high nuclear size/pleomorphism, when compared with normal epithelial cells score, 1, 2 and 3, respectively. For mitotic counts, a score is given according to the field area of the high power lens used. Only nuclei with clear features of metaphase, anaphase or telophase are included; apoptotic and hyperchromatic nuclei are discounted. Overall tumour grade is based on the final score of these three features. A score of 3, 4 or 5 implies a grade 1 cancer, a score of 6 or 7 a grade 2 tumour and 8 or 9 a carcinoma of histological grade 3. The system is easy to use and can be performed rapidly with experience. Histological grade, assessed in this way, is a strong indicator of patient survival. Patients with grade 1 carcinomas have an 85% chance of surviving for 10 years compared with less than 45% for patients with grade 3 tumours. This recognition of histological grade as an important prognostic factor has now led to methods being devised for assessing grade on cytological preparations, with the aim of obtaining prognostic information preoperatively.56 In our routine practice in Nottingham histological grade remains the factor of greatest importance in predicting prognosis for patients with primary operable breast carcinomas, with lymph node stage second in importance in multivariate analysis. All excised nodes should be examined histologically in a separate cassette. We advise that the blocks are taken by serial slicing perpendicular to the long axis of the node, thus providing the microscopist with a large area of lymph node and peripheral sinus in one section. One slice of node is examined microscopically for approximately each 5 mm of the maximum dimension of the lymph node. Lymph nodes containing any foci of carcinoma by this method are regarded as positive for staging purposes. Stage 1 tumours are lymph node negative, stage 2 tumours have three or less involved lymph nodes and stage 3 carcinomas have spread to four or more axillary and/or an internal mammary lymph node. The importance of micrometastases in prognosis remains unclear, partly because of differences in the definition ofthe size/volume and methods of assessing these lesions. Some groups have suggested that there is a prognostic association with disease free and overall survival for micrometastases identified by serial sectioning of lymph nodes,7 whereas others have found that micrometastases identified by immunohistochemical examination predict recurrence in ductal but not in lobular carcinomas.8 Although undoubtedly a small number of cases with micrometastases can be identified by immunohistochemical examination with anti-cytokeratin antibodies, we have found no prognostic significance in this feature and do Department of Histopathology, The City Hospital NHS Trust, Hucknall Road, Nottingham NG5 1PB